Build failure from CVS on OS X

Build failure from CVS on OS X

[Willy Lee]

Havne't built from source in ages, decided to start again.  Updated
with 'cvs up -PAd', reran 'configure --with-ns', kick off build with
'make bootstrap' -- build fails on nsfont.m:

gcc -I/sw/include -L/sw/lib -c  -Demacs -DHAVE_CONFIG_H
-I. -I/Users/willy/install/emacs-anonpserver/emacs/src -Dtemacs    -g
-O2 -Wno-pointer-sign  nsfont.m
nsfont.m: In function 'nsfont_draw':
nsfont.m:907: error: 'struct glyph_string' has no member named 'gidx'
nsfont.m:908: error: 'struct glyph_string' has no member named 'gidx'
nsfont.m:914: error: 'struct glyph_string' has no member named 'gidx'
nsfont.m:1097: error: 'struct glyph_string' has no member named 'gidx'
nsfont.m:1103: error: 'struct glyph_string' has no member named 'gidx'
make[1]: *** [nsfont.o] Error 1
make: *** [src] Error 2

Poking around, gidx was removed from struct glyph_string on
29.AUG.2008 by Handa-san: 'cvs log src/dispextern.h' shows thhs:

revision 1.257
date: 2008-08-29 07:54:09 +0000;  author: handa;  state: Exp;  lines:
+48 -16;  commiti
d: OnFpfDRMcaKUKEgt;
* dispextern.h (struct glyph): New union u.cmp.  Delete the member
cmp_id.
(struct glyph_string): Delete the member gidx.  New members
cmp_id, cmp_from, and cmp_to.
(enum it_method): Delete GET_FROM_COMPOSITION.
(struct composition_it): New struct.
(struct it): New member cmp_it, and iterator_stack_entry.cmp_it.
Delete c, len, cmp_id, cmp_len in u.comp.

I'm surprised noone's mentioned this before now?  I thought several
people built regularly for OS X. 

=wl 

Re: Build failure from CVS on OS X

[Randal L. Schwartz]

>>>>> "Willy" == Willy Lee <wombat@zork.net> writes:

Willy> I'm surprised noone's mentioned this before now?  I thought several
Willy> people built regularly for OS X. 

I tried building on OSX for the first time in about three months,
saw this two days ago, and reported it.  Waiting for someone to notice. :)

-- 
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<merlyn@stonehenge.com> <URL:http://www.stonehenge.com/merlyn/>
Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc.
See http://methodsandmessages.vox.com/ for Smalltalk and Seaside discussion

re: Build failure from CVS on OS X

[Clifford Wulfman]

I've posted this bug several times to bug-gnu-emacs since August 29th;  
I assumed everyone was taken up with the 22.3 release. (As I'm not a  
developer, I hesitated to post directly to emacs-devel; I'll do so in  
the future.)

--C

> >>>>> "Willy" == Willy Lee <address@hidden> writes:
>
> Willy> I'm surprised noone's mentioned this before now?  I thought  
> several
> Willy> people built regularly for OS X.
>
> I tried building on OSX for the first time in about three months,
> saw this two days ago, and reported it.  Waiting for someone to  
> notice. :)



Clifford E. Wulfman
Coordinator of Library Digital Initiatives
Princeton University Library
cwulfman@Princeton.EDU

Re: Build failure from CVS on OS X

[Randal L. Schwartz]

>>>>> "Clifford" == Clifford Wulfman <cwulfman@Princeton.EDU> writes:

Clifford> I've posted this bug several times to bug-gnu-emacs since August
Clifford> 29th; I assumed everyone was taken up with the 22.3 release. (As I'm
Clifford> not a developer, I hesitated to post directly to emacs-devel; I'll
Clifford> do so in the future.)

Well, something got fixed, because the build I just did, worked.

Now to figure out why my font settings from my 22.1 release don't
work with 23.x, and how many other things are broken.

-- 
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<merlyn@stonehenge.com> <URL:http://www.stonehenge.com/merlyn/>
Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc.
See http://methodsandmessages.vox.com/ for Smalltalk and Seaside discussion

Re: Build failure from CVS on OS X

[Randal L. Schwartz]

>>>>> "Randal" == Randal L Schwartz <merlyn@stonehenge.com> writes:

Randal> Now to figure out why my font settings from my 22.1 release don't
Randal> work with 23.x, and how many other things are broken.

I'm now live on 23.0.60 CVS HEAD on OSX.  First new behavior change (can
someone explain this or tell me how to fix it):

If I'm in a shell buffer inside the sole window of my Emacs, and invoke
"emacsclient" indirectly through some git commit, the 22.1 behavior replaced
my shell window with the emacs window, waited for my M-#, then returned me to
my shell window.  The 23.x behavior seems to be "return me to a random other
buffer".  Oops.

-- 
Randal L. Schwartz - Stonehenge Consulting Services, Inc. - +1 503 777 0095
<merlyn@stonehenge.com> <URL:http://www.stonehenge.com/merlyn/>
Smalltalk/Perl/Unix consulting, Technical writing, Comedy, etc. etc.
See http://methodsandmessages.vox.com/ for Smalltalk and Seaside discussion

Re: Build failure from CVS on OS X

[Ted Zlatanov]

On Sun, 07 Sep 2008 11:42:20 -0700 merlyn@stonehenge.com (Randal L. Schwartz) wrote: 

>>>>>> "Clifford" == Clifford Wulfman <cwulfman@Princeton.EDU> writes:
Clifford> I've posted this bug several times to bug-gnu-emacs since August
Clifford> 29th; I assumed everyone was taken up with the 22.3 release. (As I'm
Clifford> not a developer, I hesitated to post directly to emacs-devel; I'll
Clifford> do so in the future.)

RLS> Well, something got fixed, because the build I just did, worked.

Just look in src/ChangeLog.  I made the change; it's probably
insufficient and composition is known to be broken, but at least the
compilation can proceed.

RLS> Now to figure out why my font settings from my 22.1 release don't
RLS> work with 23.x, and how many other things are broken.

I've seen random crashes, but have not been able to track them down.
Keep an eye out for them.

I've also reported a read-passwd bug: you have to use C-m, not RET to
finish input.  RET makes a new line in the minibuffer.  Very confusing
to new users.

Ted

Re: Build failure from CVS on OS X

[Robert Casties]

Ted Zlatanov wrote:
> I've also reported a read-passwd bug: you have to use C-m, not RET to
> finish input.  RET makes a new line in the minibuffer.  Very confusing
> to new users.

Do you still have that problem? It went away for me already weeks ago 
after a fix in ns-win.el (18. August).

Cheers
	Robert Casties

Re: Build failure from CVS on OS X

[Ted Zlatanov]

On Wed, 10 Sep 2008 18:10:08 +0200 Robert Casties <casties@mpiwg-berlin.mpg.de> wrote: 

RC> Ted Zlatanov wrote:
>> I've also reported a read-passwd bug: you have to use C-m, not RET to
>> finish input.  RET makes a new line in the minibuffer.  Very confusing
>> to new users.

RC> Do you still have that problem? It went away for me already weeks ago
RC> after a fix in ns-win.el (18. August).

I tested last night and it's no longer a problem.

Thanks
Ted